SERC Supports U.S. Navy’s Military Sealift Command in Development and Use of Digital Twins
In 2018, the U.S. Department of Defense (DoD) delivered their Digital Engineering Strategy to transform traditional engineering practice into digital engineering. Digital engineering will enable the DoD to rapidly infuse advanced technologies, such as Artificial Intelligence (AI), Machine Learning (ML), Internet of Things (IoT), Digital Twin (DT), 3D printing, etc., into acquisitions. The U.S. Navy’s Military Sealift Command (MSC) is actively employing digital engineering techniques within their ship maintenance division. MSC is the premier provider of ocean transportation to the DoD. The Command operates approximately 125 civilian-crewed ships that replenish U.S. Navy ships, conduct specialized missions, strategically pre-position combat cargo at sea around the world and move military cargo and supplies used by deployed U.S. forces and coalition partners.
MSC has invested considerable capital and resources in responding to the digital engineering strategy initiative. Initial work is the ongoing development of a digital model – the virtual representation of the physical ship– that will be used to improve analysis of ship components and wear. This information will be used to increase ship availability and improve operations. Additionally, MSC is building towards a full digital twin (DT), the virtual representation of the physical ship and the data flows from the existing physical object that are fully integrated in both directions, that will be used to improve their current predictive maintenance program. Ultimately the goal is to provide an environment where operations and human activities with the twin can be modeled and considered.
Principal Investigator: Dr. Sam Kovacic, Old Dominion University
Dr. Sam Kovacic, Old Dominion University, the Principal Investigator for this task, acknowledges, “Although the benefits of Digital Engineering are significant, its greater contribution is its integration into the random nature of real-world operations.” Currently, a digital twin is still a closed system with respect to the environment and operations. Kovacic adds, “Expanding the DT into an open system within an environment that can accept data input from real world activities will serve MSC’s vision of a reliability centered maintenance strategy that enables understanding the performance of their ships beyond an instance in time and under variable conditions.”
To facilitate this vision, the SERC research team is working with MSC on two concurrent activities: first, exploring methods to consume current data in all its forms into data to create a digital model, and developing ways to create the linkages between the model and the ship itself to build the digital twin. These efforts will create the predictive capability necessary for MSC to maintain the ship as a whole rather than its constituent parts. The second activity is to define a framework for an environment that allows for a decision space where the result of the model and the data within the domain can be consumed by the participants to forecast and execute a maintenance strategy that can address the random nature of the real world.
This research focuses on the understanding necessary to address a problem set that cannot be addressed with a traditional systems approach. This strategy places design, complexity and understanding with the perspective rather than with the artifact. Within this paradigm complexity is not an attribute of an artifact, it is a condition that arises when there is not the requisite degree of understanding to solve a problem given an intended approach. By focusing on the observer’s perspective rather than the artifact and its implication and/or influences, it is possible to zero in on the nature of understanding in a situation made complex by multiple perspectives, change, and the dynamics of the solution. The aim is to advance techniques that improve design and decision making and how to proceed when a high level of complexity is present, and a requisite level of understanding cannot be achieved.
This approach would not be used necessarily for designing a new system, but would provide the ability to forecast, develop prognoses, and generate understanding in a stochastic domain that includes the system, i.e., a human compatible digital twin, within an environment that explores the random nature of the domain, rather than the predictive nature of the artifact. It is anticipated that this will lead to better understanding and improve the decision process by making the digital twin compatible with the human generative process of the decision maker that must make strategic decisions within this complex domain.